A104: Thromboembolism in patients with localized upper Gastro-intestinal cancers undergoing radical treatment.

Abdul Qayam1,Helen Wong1,Helen Neville-Webbe1

1Clatterbridge Cancer Centre, Merseyside, UK,2Clatterbridge Cancer Centre, Merseyside, UK,3Clatterbridge Cancer Centre, Merseyside, UK

Presenting date: Monday 2 November
Presenting time: 13.10-14.00


Thromboembolism (TE) is relatively common in cancer patients. It is more common in those with advanced disease and those who are receiving chemotherapy. The type of chemotherapy also affects the risk. This present study was undertaken to explore the incidence of TE in patients with upper gastro-intestinal cancers being treated with curative intent. Current national guidelines does not recommend prophylactic anticoagulation for this cancer group.



Data was collected retrospectively for 387 patients treated at Clatterbridge Cancer Centre from March 2009 to October 2013. Only the patients with confirmed histological diagnosis and properly worked-up staging were included. Those who were on anti-coagulation treatment, whether therapeutic or prophylactic, at the time of start of chemotherapy were excluded. Data was stratified according to tumour site: oesophagus (O), stomach (S) or gastro-oesophageal junction (GOJ), and according to chemotherapy received: cisplatin/5-Fluorouracil (cis/5FU) , epirubicin, cisplatin and capecitabine (ECX), cisplatin & capecitabine (cis/cape), and ‘various' (includes epirubicin alone, ECF, bevacizumab with ECX, capecitabine or 5-FU alone). Co-morbidity status on initial presentation was examined, with empirical scoring for co-morbidity status used. The patients were stratified into following groups: Nil (‘no co-morbidity'), mild (‘co-morbidities without end-organ damage and no need for dose modification'), moderate (‘co-morbidity has effect on end-organs, need careful dosing'), or severe. The last group was excluded as they did not receive any chemotherapy treatment.


73.8% (n=279) were adenocarcinoma, and 22.8% (n=86) were squamous cell carcinoma. The mean age was 64.8±9.3 years. 261 (69%) were male and 117 (31 %) female. The primary tumour was in the O, S or GOJ in 60.3%, 22.8% and 16.9%, respectively. 52 patients (13.4%) had a TE event. The incidence was 11.1% if primary tumour was located in the O, as opposed to 20% and 18% if S or GOJ respectively. TE occurred in 4.5% of patients receiving  cisp/5FU, 11.5% receiving cis/cape, but in 21.3% receiving ECX. TE occurred in 5.3% of the patients who recieved 5FU, compared to16.1% in capecitabine group (p=0.15). A total of 54 (14%) patients received also radiotherapy (concurrent or sequential). The incidence of TE in this group was 14.8% (n=8/54), and not significantly different from the incidence in patients not treated with radiotherapy (p=0.91). Co-morbidty did not significantly affect risk of TE. Venous TE was more common than arterial (69.2% versus 30.8%).


The incidence of TE in patients with upper GI cancers undergoing radical treatment is a serious concern. The incidence of 13.4% is comparable to that reported in other studies. However we did find that the incidence was significantly higher in patients receiving capecitabine. The treatment had to be discontinued in 5.7% of cases, and needed delays in treatment in further 38.4% of the patients.  Most of the TE events occurred during first two months of diagnosis. The occurrence of TE did not affect 1-year OS. The present study does not support any deviation from TE prophylaxis guidelines by ASCO, NCCN and ESMO